New Pathways in Biotech Innovation
Students tackle case studies on stem cells and IBM’s use of AI in precision medicine
What if you could help save 20,000 lives every year with a new cancer drug?
Or you could provide one more year for a patient to spend with her grandchild, as Biotechnology Industry Immersion coordinator Judith Kjelstrom puts it.
Targeted medicine is the topic of the day for the MBA, doctoral and other graduate students as they team up on live case studies from two industry executives Kjelstrom has invited.
Christopher Murriel of OncoMed Pharmaceuticals poses a real challenge from the pharma company. Students have one hour to form their strategies. What they present could influence how Murriel and his company proceed through this actual situation:
Your lab has identified the genetic markers for an aggressive form of cancer. It often spreads quickly to the brain, bone, kidney and other organs. The lab has developed a model of how specific receptors within a malignant tumor’s stem cells, when turned off, can prevent the cancer from growing. Your company is ready to invest hundreds of millions of dollars into a new drug based on this data. How would you incorporate this information into your strategy for the clinical trials? What may be some unforeseen consequences? What do you do if it fails?
The students grasp for more details: Are there side effects? What is the quality of life? Why add another drug for these late-stage patients when nothing else has worked?
“It’s a fantastic opportunity for students to get some knowledge and classroom learning about the intersection between business, academia and biotechnology,” says Murriel.
“It’s especially a great opportunity for those who don’t have MBAs but are in Ph.D. programs to learn from those who are.”
Big Data in Genomics and healthcare
The promise of personalized medicine has been on the horizon for some time. Leveraging healthcare big data in a meaningful way is key to making it happen. IBM is on the forefront.
IBM’s Vice President of Research Jeff Welser explains to students how his lab is applying IBM’s machine learning tools derived from the famous Watson supercomputer toward crunching big data in genomics, offering personalized treatment strategies for doctors and patients.
“It’s changing everything we think about when it comes to understanding traits,” says Welser.
“You can sequence your own genome and the specific cancer you have with the specific mutation you have and compare it to others.”
Personalized data is exploding in healthcare, he says, with far-reaching impacts for the future of medicine. “But nobody understands how these things work as well as they do,” he adds. “That means you don’t know how to improve them necessarily.”
He breaks down the process. When using artificial intelligence to identify cats in photos, for instance, you first show it cat pictures. The neural network runs through a series of algorithms and identifies key traits, like pointy ears. If it then correctly guesses which one is a cat in the images, you reward it by leaving that arrangement of nodes on and then refining the process from there.
But you never know exactly what traits it’s identifying. You may learn it’s looking at pointy ears when it labels a dog with pointy ears as a cat. Or you may never discover how the system came to that conclusion. It can lead to tricky ethical issues when the AI is sifting through 10,000 MRI scans to decide, in context with a patient’s medical records, whether a lump is a melanoma or is benign.
Bringing AI to life
Welser weaves in other examples of challenges and opportunities in AI: hacking self-driving cars, deciding who gets loans, identifying pathogens by sequencing snippets of DNA from food bacteria in a factory, and tracking subtle indications of Parkinson’s disease in a patient’s movements by outfitting her house with sensors.
He spins out scenarios that sound like plot devices from sci-fi novels.
His live case study for the students examines the over-prescription of antibiotics in healthcare, which intersects biotech with business, food, agriculture and sustainability.
In 10 to 15 years, we’ll look back on our use of vitamins as being “cute.” But what we’re really interested in is targeting ways to specifically improve our own microbiomes.